GETGRAIL · EVIDENCE LAB

Gold Trading Research

Decision support for XAU/USD. Historical evidence, explicit uncertainty, no automated execution.

RESEARCH ONLY
LIVE DECISION SUPPORTLoad a validated real-data report to enable continuous monitoring.
WhatsApp: NOT CONFIGURED Last checked not yet

DATA INGEST

Load licensed one-minute XAU/USD data

Required columns: timestamp, open, high, low, close. UTC offsets and genuine session gaps are preserved.

REAL-DATA CONNECTOR DISABLED Twelve Data /time_series · API key not configured
Synthetic sample loaded; market conclusions remain disabled.
SYNTHETIC SAMPLE — NO MARKET CONCLUSION Local deterministic synthetic OHLC sample v1 · 721 candles · 2026-01-05 08:00:00Z → 2026-01-05 20:00:00Z
REPRODUCIBLE EXPORTExperiment 7DC0F5E48C677296AA29 · source and decision gates included
Download audit JSON Download evidence CSV

COMPARABLE SITUATIONS

Current regime: Range

±0.5% · 30-minute outcome
ImpulseRegimeNHit rate95% intervalMedian directionalAdverse excursion
DownRange1+100.00%+20.65%–+100.00%+0.28%+0.03%

CURRENT READOUT

NO TRADE

Synthetic data can never qualify a research scenario.

Price
1999.00
15m move
-0.01%
Regime
Range
Session
UnitedStates
Confidence
Dataset ineligible for current setup

SCENARIO LEVELS

Withheld

No entry, invalidation, stop, or target is shown while evidence is weak or no setup is active.

DECISION AUDIT

Why this readout is allowed or blocked

0/11 gates passed
BLOCK

Authentic market data

Synthetic fixture; market qualification prohibited.

BLOCK

Historical coverage

12.0 hours; minimum 48.

BLOCK

Current-data freshness

Latest candle age 6787.9 hours; maximum 24.

BLOCK

Active ±0.5% / 15m setup

Observed -0.01%.

BLOCK

Comparable evidence count

0/30 comparable events.

BLOCK

Net edge and uncertainty

No comparable regime evidence.

BLOCK

Historical reward/risk

Median net move / median adverse excursion = 0.00; minimum 1.00.

BLOCK

Multiple-testing control

Bonferroni-adjusted one-sided p = 1.0000; family-wise limit 0.0500.

BLOCK

Chronological holdout

No direction-specific holdout available.

BLOCK

Walk-forward replication

No direction-specific walk-forward result available.

BLOCK

Threshold stability

No direction-specific stability result available.

DATA QUALITY

CONTINUOUS

Exact one-minute cadence expected
Coverage
+100.00%
Gap count
0
Missing minutes
0
Largest gap
0 min

FIRST EXPERIMENT

After a ±0.5% move within 15 minutes

Continuation lens
ImpulseHorizonNNet hit rate95% intervalAvg next moveMedian net edgeAvg adverseVolatility
Up 5 min4+50.00% +15.00%–+85.00% +0.11%+0.09% +0.03%+0.02%
Up 15 min4+50.00% +15.00%–+85.00% 0.00%-0.03% +0.08%+0.02%
Up 30 min40.00% 0.00%–+48.99% -0.34%-0.37% +0.36%+0.02%
Up 60 min40.00% 0.00%–+48.99% -0.58%-0.59% +0.61%+0.02%
Up 120 min40.00% 0.00%–+48.99% -0.53%-0.55% +1.16%+0.02%
Down 5 min5+60.00% +23.07%–+88.24% -0.08%+0.04% +0.03%+0.02%
Down 15 min5+60.00% +23.07%–+88.24% +0.02%+0.03% +0.10%+0.02%
Down 30 min5+20.00% +3.62%–+62.45% +0.23%-0.25% +0.31%+0.02%
Down 60 min5+20.00% +3.62%–+62.45% +0.45%-0.51% +0.51%+0.02%
Down 120 min50.00% 0.00%–+43.45% +0.54%-0.54% +1.05%+0.02%

LONG VS SHORT

Continuation compared with mean reversion

±0.5% impulse · costs included
ImpulseHorizonNContinuationNet hitAvg net moveReversalNet hitAvg net move
Up5 min4LONG+50.00%+0.09%SHORT+25.00%-0.13%
Up15 min4LONG+50.00%-0.02%SHORT+50.00%-0.02%
Up30 min4LONG0.00%-0.36%SHORT+100.00%+0.32%
Up60 min4LONG0.00%-0.60%SHORT+100.00%+0.56%
Up120 min4LONG0.00%-0.55%SHORT+100.00%+0.51%
Down5 min5SHORT+60.00%+0.06%LONG+20.00%-0.10%
Down15 min5SHORT+60.00%-0.04%LONG+40.00%0.00%
Down30 min5SHORT+20.00%-0.25%LONG+80.00%+0.21%
Down60 min5SHORT+20.00%-0.47%LONG+80.00%+0.43%
Down120 min5SHORT0.00%-0.56%LONG+100.00%+0.52%
This comparison is descriptive. A reversal is only research-qualified when every separate validation gate below passes.

REVERSAL VALIDATION

Mean-reversion evidence tested separately

±0.5% impulse · 30m · costs included
ImpulseResearch sideNNet hitWilson lowAvg netHoldout hitWF foldsStable thresholdsAdjusted pR/RResult
UpSHORT4+100.00%+51.01%+0.32%+100.00%0/301.00001.98NO TRADE
DownLONG5+80.00%+37.55%+0.21%+100.00%0/301.00001.24NO TRADE
No automatic execution. A descriptive reversal advantage remains NO TRADE when holdout, forward replication, uncertainty, stability, multiple-testing, or reward/risk fails.

COST STRESS

Does the continuation survive friction?

0× · configured · 2× round-trip cost
ImpulseHorizonRound-trip costNNet hit rateAverage netMedian net
Up5 min0.00%4+75.00%+0.11%+0.11%
Up5 min+0.02%4+50.00%+0.09%+0.09%
Up5 min+0.04%4+50.00%+0.07%+0.07%
Up15 min0.00%4+50.00%0.00%-0.01%
Up15 min+0.02%4+50.00%-0.02%-0.03%
Up15 min+0.04%4+50.00%-0.04%-0.05%
Up30 min0.00%40.00%-0.34%-0.35%
Up30 min+0.02%40.00%-0.36%-0.37%
Up30 min+0.04%40.00%-0.38%-0.39%
Up60 min0.00%40.00%-0.58%-0.57%
Up60 min+0.02%40.00%-0.60%-0.59%
Up60 min+0.04%40.00%-0.62%-0.61%
Up120 min0.00%40.00%-0.53%-0.53%
Up120 min+0.02%40.00%-0.55%-0.55%
Up120 min+0.04%40.00%-0.57%-0.57%
Down5 min0.00%5+60.00%+0.08%+0.06%
Down5 min+0.02%5+60.00%+0.06%+0.04%
Down5 min+0.04%5+60.00%+0.04%+0.02%
Down15 min0.00%5+60.00%-0.02%+0.05%
Down15 min+0.02%5+60.00%-0.04%+0.03%
Down15 min+0.04%5+60.00%-0.06%+0.01%
Down30 min0.00%5+20.00%-0.23%-0.23%
Down30 min+0.02%5+20.00%-0.25%-0.25%
Down30 min+0.04%5+20.00%-0.27%-0.27%
Down60 min0.00%5+20.00%-0.45%-0.49%
Down60 min+0.02%5+20.00%-0.47%-0.51%
Down60 min+0.04%5+20.00%-0.49%-0.53%
Down120 min0.00%50.00%-0.54%-0.52%
Down120 min+0.02%50.00%-0.56%-0.54%
Down120 min+0.04%50.00%-0.58%-0.56%
This is a parameter stress test, not a bid/ask reconstruction. Use feed-level bid/ask data before execution-oriented interpretation.

SESSION COMPARABLES

Current UTC session: UnitedStates

±0.5% · 30-minute net outcome
ImpulseSessionNNet hit rate95% intervalMedian net edge
DownUnitedStates10.00%0.00%–+79.35%-0.53%

CHRONOLOGICAL HOLDOUT

Development 70% · untouched holdout 30%

Estimated round-trip cost: +0.02%
DirectionDevelopment NDevelopment net hitHoldout NHoldout net hitHoldout median netResult
Up20.00%20.00%-0.20%NO TRADE
Down3+33.33%20.00%-0.34%NO TRADE

WALK-FORWARD

Three sequential forward test blocks

Expanding training · ±0.5% · 30m
DirectionFoldTrain NTrain hitTest NTest hitTest medianResult
Up120.00%00.00%0.00%NO TRADE
Up220.00%10.00%-0.12%NO TRADE
Up330.00%10.00%-0.28%NO TRADE
Down12+50.00%10.00%-0.25%NO TRADE
Down23+33.33%10.00%-0.15%NO TRADE
Down34+25.00%10.00%-0.53%NO TRADE

INDEPENDENCE CONTROL

15-minute event cooldown

Repeated crossings in the same direction and threshold inside the cooldown are excluded to reduce overlapping, pseudo-independent observations.

PARAMETER STABILITY

Neighboring impulse thresholds

30-minute net continuation
DirectionQualified thresholdsAvailable thresholdsResultConclusion
Up02NO TRADENet continuation is not stable across at least two sufficiently populated thresholds.
Down02NO TRADENet continuation is not stable across at least two sufficiently populated thresholds.

RESEARCH UNIVERSE

±+0.25%20 detected events with a 30m outcome
±+0.50%9 detected events with a 30m outcome
±+1.00%0 detected events with a 30m outcome

• Research/decision support only; no automatic execution and no guaranteed signal.

• LOCAL SYNTHETIC SAMPLE: validates calculations only; it cannot support conclusions about XAU/USD behavior.

• Current-setup gate: Synthetic data can never qualify a research scenario.

• Evidence is below the minimum decision threshold; scenario levels are intentionally withheld.

Method: new threshold crossings only; exact timestamp horizons; gaps are never bridged; 95% hit-rate uncertainty uses a Wilson interval.

Experiment 7DC0F5E48C677296AA29 · gold-research-v5-reversal-validation · window 15 min · thresholds 0.25 %, 0.50 %, 1.00 % · horizons 5m, 15m, 30m, 60m, 120m · minimum N 30 · minimum hit +58.00% · minimum R/R 1.00 · family-wise α 0.050 · holdout +30.00% · history ≥ 48 h · current-data age ≤ 24 h